[PATCH] change CONFIG_NUMA description

[PATCH] change CONFIG_NUMA description

[KOSAKI Motohiro]

CONFIG_NUMA description talk about a bit old thing.
So, following changes are better.

o CONFIG_NUMA is no longer EXPERIMENTAL
o Opteron is not the only processor of NUMA topology on x86_64 no longer, but also Intel Core7i has it.


Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
---
 arch/x86/Kconfig |    7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

Index: b/arch/x86/Kconfig
===================================================================
--- a/arch/x86/Kconfig	2008-11-05 01:11:24.000000000 +0900
+++ b/arch/x86/Kconfig	2008-11-05 02:09:38.000000000 +0900
@@ -951,7 +951,7 @@ config ARCH_PHYS_ADDR_T_64BIT
 
 # Common NUMA Features
 config NUMA
-	bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
+	bool "Numa Memory Allocation and Scheduler Support"
 	depends on SMP
 	depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)
 	default n if X86_PC
@@ -964,9 +964,8 @@ config NUMA
 
 	  For 32-bit this is currently highly experimental and should be only
 	  used for kernel development. It might also cause boot failures.
-	  For 64-bit this is recommended on all multiprocessor Opteron systems.
-	  If the system is EM64T, you should say N unless your system is
-	  EM64T NUMA.
+	  For 64-bit this is recommended if the system is Intel Core 7i or later or AMD Opteron.
+	  Otherwise, you should say N unless your system is EM64T NUMA.
 
 comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
 	depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI)

Re: [PATCH] change CONFIG_NUMA description

[Chris Snook]

KOSAKI Motohiro wrote:
> CONFIG_NUMA description talk about a bit old thing.
> So, following changes are better.
> 
> o CONFIG_NUMA is no longer EXPERIMENTAL
> o Opteron is not the only processor of NUMA topology on x86_64 no longer, but also Intel Core7i has it.
> 
> 
> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>

a) It's Core i7, not Core 7i
b) Core i7 is, at least so far, the single-socket version, so NUMA 
doesn't apply.

I think we should let Intel write the CONFIG_NUMA description for 
Nehalem processors.

-- Chris

Re: [PATCH] change CONFIG_NUMA description

[Pekka Enberg]

On Tue, Nov 4, 2008 at 7:27 PM, KOSAKI Motohiro
<kosaki.motohiro@jp.fujitsu.com> wrote:
> CONFIG_NUMA description talk about a bit old thing.
> So, following changes are better.
>
> o CONFIG_NUMA is no longer EXPERIMENTAL
> Index: b/arch/x86/Kconfig
> ===================================================================
> --- a/arch/x86/Kconfig  2008-11-05 01:11:24.000000000 +0900
> +++ b/arch/x86/Kconfig  2008-11-05 02:09:38.000000000 +0900
> @@ -951,7 +951,7 @@ config ARCH_PHYS_ADDR_T_64BIT
>
>  # Common NUMA Features
>  config NUMA
> -       bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
> +       bool "Numa Memory Allocation and Scheduler Support"
>        depends on SMP
>        depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)

You might want to remove the dependency to CONFIG_EXPERIMENTAL as well.

>        default n if X86_PC
> @@ -964,9 +964,8 @@ config NUMA
>
>          For 32-bit this is currently highly experimental and should be only
>          used for kernel development. It might also cause boot failures.
> -         For 64-bit this is recommended on all multiprocessor Opteron systems.
> -         If the system is EM64T, you should say N unless your system is
> -         EM64T NUMA.
> +         For 64-bit this is recommended if the system is Intel Core 7i or later or AMD Opteron.
> +         Otherwise, you should say N unless your system is EM64T NUMA.
>
>  comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
>        depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI)
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Re: [PATCH] change CONFIG_NUMA description

[Ingo Molnar]

* KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:

> 
> CONFIG_NUMA description talk about a bit old thing.
> So, following changes are better.
> 
> o CONFIG_NUMA is no longer EXPERIMENTAL o Opteron is not the only 
> processor of NUMA topology on x86_64 no longer, but also Intel 
> Core7i has it.
> 
> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> ---
>  arch/x86/Kconfig |    7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)

applied to tip/x86/doc, thanks! I've updated a few other details as 
well, see the final commit below.

	Ingo

------------->
>From fd51b2d7d5df932767b89e00d0871a38a2c53e74 Mon Sep 17 00:00:00 2001
From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Date: Wed, 5 Nov 2008 02:27:19 +0900
Subject: [PATCH] x86: update CONFIG_NUMA description

Impact: clarify/update CONFIG_NUMA text

CONFIG_NUMA description talk about a bit old thing.
So, following changes are better.

 o CONFIG_NUMA is no longer EXPERIMENTAL

 o Opteron is not the only processor of NUMA topology on x86_64 no longer,
   but also Intel Core7i has it.

Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/Kconfig |   16 ++++++++++------
 1 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 350bee1..38ae04b 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -951,22 +951,26 @@ config ARCH_PHYS_ADDR_T_64BIT
 
 # Common NUMA Features
 config NUMA
-	bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
+	bool "Numa Memory Allocation and Scheduler Support"
 	depends on SMP
 	depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)
 	default n if X86_PC
 	default y if (X86_NUMAQ || X86_SUMMIT || X86_BIGSMP)
 	help
 	  Enable NUMA (Non Uniform Memory Access) support.
+
 	  The kernel will try to allocate memory used by a CPU on the
 	  local memory controller of the CPU and add some more
 	  NUMA awareness to the kernel.
 
-	  For 32-bit this is currently highly experimental and should be only
-	  used for kernel development. It might also cause boot failures.
-	  For 64-bit this is recommended on all multiprocessor Opteron systems.
-	  If the system is EM64T, you should say N unless your system is
-	  EM64T NUMA.
+	  For 64-bit this is recommended if the system is Intel Core 7i
+	  (or later), AMD Opteron, or EM64T NUMA.
+
+	  For 32-bit this is only needed on (rare) 32-bit-only platforms
+	  that support NUMA topologies, such as NUMAQ / Summit, or if you
+	  boot a 32-bit kernel on a 64-bit NUMA platform.
+
+	  Otherwise, you should say N.
 
 comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
 	depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI)

Re: [PATCH] change CONFIG_NUMA description

[Pekka Enberg]

Hi Ingo,

On Thu, Nov 6, 2008 at 10:53 AM, Ingo Molnar <mingo@elte.hu> wrote:
> From fd51b2d7d5df932767b89e00d0871a38a2c53e74 Mon Sep 17 00:00:00 2001
> From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> Date: Wed, 5 Nov 2008 02:27:19 +0900
> Subject: [PATCH] x86: update CONFIG_NUMA description
>
> Impact: clarify/update CONFIG_NUMA text
>
> CONFIG_NUMA description talk about a bit old thing.
> So, following changes are better.
>
>  o CONFIG_NUMA is no longer EXPERIMENTAL
>
>  o Opteron is not the only processor of NUMA topology on x86_64 no longer,
>   but also Intel Core7i has it.
>
> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> ---
>  arch/x86/Kconfig |   16 ++++++++++------
>  1 files changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index 350bee1..38ae04b 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -951,22 +951,26 @@ config ARCH_PHYS_ADDR_T_64BIT
>
>  # Common NUMA Features
>  config NUMA
> -       bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
> +       bool "Numa Memory Allocation and Scheduler Support"
>        depends on SMP
>        depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)

Why does this depend on CONFIG_EXPERIMENTAL still?

>        default n if X86_PC
>        default y if (X86_NUMAQ || X86_SUMMIT || X86_BIGSMP)
>        help
>          Enable NUMA (Non Uniform Memory Access) support.
> +
>          The kernel will try to allocate memory used by a CPU on the
>          local memory controller of the CPU and add some more
>          NUMA awareness to the kernel.
>

Re: [PATCH] change CONFIG_NUMA description

[KOSAKI Motohiro]

> Hi Ingo,
> 
> On Thu, Nov 6, 2008 at 10:53 AM, Ingo Molnar <mingo@elte.hu> wrote:
> > From fd51b2d7d5df932767b89e00d0871a38a2c53e74 Mon Sep 17 00:00:00 2001
> > From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> > Date: Wed, 5 Nov 2008 02:27:19 +0900
> > Subject: [PATCH] x86: update CONFIG_NUMA description
> >
> > Impact: clarify/update CONFIG_NUMA text
> >
> > CONFIG_NUMA description talk about a bit old thing.
> > So, following changes are better.
> >
> >  o CONFIG_NUMA is no longer EXPERIMENTAL
> >
> >  o Opteron is not the only processor of NUMA topology on x86_64 no longer,
> >   but also Intel Core7i has it.
> >
> > Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> > Signed-off-by: Ingo Molnar <mingo@elte.hu>
> > ---
> >  arch/x86/Kconfig |   16 ++++++++++------
> >  1 files changed, 10 insertions(+), 6 deletions(-)
> >
> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > index 350bee1..38ae04b 100644
> > --- a/arch/x86/Kconfig
> > +++ b/arch/x86/Kconfig
> > @@ -951,22 +951,26 @@ config ARCH_PHYS_ADDR_T_64BIT
> >
> >  # Common NUMA Features
> >  config NUMA
> > -       bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
> > +       bool "Numa Memory Allocation and Scheduler Support"
> >        depends on SMP
> >        depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)
> 
> Why does this depend on CONFIG_EXPERIMENTAL still?

sorry, late resoponce.
In fact, x86_32 numa is still experimental.
and x86_64 is already NOT experimental.

So, I think depend-on don't need change.

Re: [PATCH] change CONFIG_NUMA description

[Ingo Molnar]

* Pekka Enberg <penberg@cs.helsinki.fi> wrote:

> Hi Ingo,
> 
> On Thu, Nov 6, 2008 at 10:53 AM, Ingo Molnar <mingo@elte.hu> wrote:
> > From fd51b2d7d5df932767b89e00d0871a38a2c53e74 Mon Sep 17 00:00:00 2001
> > From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> > Date: Wed, 5 Nov 2008 02:27:19 +0900
> > Subject: [PATCH] x86: update CONFIG_NUMA description
> >
> > Impact: clarify/update CONFIG_NUMA text
> >
> > CONFIG_NUMA description talk about a bit old thing.
> > So, following changes are better.
> >
> >  o CONFIG_NUMA is no longer EXPERIMENTAL
> >
> >  o Opteron is not the only processor of NUMA topology on x86_64 no longer,
> >   but also Intel Core7i has it.
> >
> > Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> > Signed-off-by: Ingo Molnar <mingo@elte.hu>
> > ---
> >  arch/x86/Kconfig |   16 ++++++++++------
> >  1 files changed, 10 insertions(+), 6 deletions(-)
> >
> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > index 350bee1..38ae04b 100644
> > --- a/arch/x86/Kconfig
> > +++ b/arch/x86/Kconfig
> > @@ -951,22 +951,26 @@ config ARCH_PHYS_ADDR_T_64BIT
> >
> >  # Common NUMA Features
> >  config NUMA
> > -       bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
> > +       bool "Numa Memory Allocation and Scheduler Support"
> >        depends on SMP
> >        depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)
> 
> Why does this depend on CONFIG_EXPERIMENTAL still?

Good point, i've just removed that too. 32-bit NUMA used to have 
stability problems (nobody really tested it systematically), but we've 
fixed those bugs and can now remove the restrictions in v2.6.29.

	Ingo

Re: [PATCH] change CONFIG_NUMA description

[Andi Kleen]

KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> writes:

> CONFIG_NUMA description talk about a bit old thing.
> So, following changes are better.
>
> o CONFIG_NUMA is no longer EXPERIMENTAL
> o Opteron is not the only processor of NUMA topology on x86_64 no longer, but also Intel Core7i has it.

Actually it doesn't have it. Core i7 systems are single socket and therefore
not NUMA.


-Andi

-- 
ak@linux.intel.com

Re: [PATCH] change CONFIG_NUMA description

[Andi Kleen]

Chris Snook <csnook@redhat.com> writes:

> KOSAKI Motohiro wrote:
>> CONFIG_NUMA description talk about a bit old thing.
>> So, following changes are better.
>> o CONFIG_NUMA is no longer EXPERIMENTAL
>> o Opteron is not the only processor of NUMA topology on x86_64 no longer, but also Intel Core7i has it.
>> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
>
> a) It's Core i7, not Core 7i
> b) Core i7 is, at least so far, the single-socket version, so NUMA
> doesn't apply.
>
> I think we should let Intel write the CONFIG_NUMA description for
> Nehalem processors.

The recommendation should be imho just to always enable it. Back 
when I wrote the description it was still experimental, but 
it really isn't anymore.

-Andi

-- 
ak@linux.intel.com

Re: [PATCH] change CONFIG_NUMA description

[Chris Snook]

Ingo Molnar wrote:
> * KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> wrote:
> 
>> CONFIG_NUMA description talk about a bit old thing.
>> So, following changes are better.
>>
>> o CONFIG_NUMA is no longer EXPERIMENTAL o Opteron is not the only 
>> processor of NUMA topology on x86_64 no longer, but also Intel 
>> Core7i has it.
>>
>> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
>> ---
>>  arch/x86/Kconfig |    7 +++----
>>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> applied to tip/x86/doc, thanks! I've updated a few other details as 
> well, see the final commit below.
> 
> 	Ingo
> 
> ------------->
> From fd51b2d7d5df932767b89e00d0871a38a2c53e74 Mon Sep 17 00:00:00 2001
> From: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> Date: Wed, 5 Nov 2008 02:27:19 +0900
> Subject: [PATCH] x86: update CONFIG_NUMA description
> 
> Impact: clarify/update CONFIG_NUMA text
> 
> CONFIG_NUMA description talk about a bit old thing.
> So, following changes are better.
> 
>  o CONFIG_NUMA is no longer EXPERIMENTAL
> 
>  o Opteron is not the only processor of NUMA topology on x86_64 no longer,
>    but also Intel Core7i has it.
> 
> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> ---
>  arch/x86/Kconfig |   16 ++++++++++------
>  1 files changed, 10 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index 350bee1..38ae04b 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -951,22 +951,26 @@ config ARCH_PHYS_ADDR_T_64BIT
>  
>  # Common NUMA Features
>  config NUMA
> -	bool "Numa Memory Allocation and Scheduler Support (EXPERIMENTAL)"
> +	bool "Numa Memory Allocation and Scheduler Support"
>  	depends on SMP
>  	depends on X86_64 || (X86_32 && HIGHMEM64G && (X86_NUMAQ || X86_BIGSMP || X86_SUMMIT && ACPI) && EXPERIMENTAL)
>  	default n if X86_PC
>  	default y if (X86_NUMAQ || X86_SUMMIT || X86_BIGSMP)
>  	help
>  	  Enable NUMA (Non Uniform Memory Access) support.
> +
>  	  The kernel will try to allocate memory used by a CPU on the
>  	  local memory controller of the CPU and add some more
>  	  NUMA awareness to the kernel.
>  
> -	  For 32-bit this is currently highly experimental and should be only
> -	  used for kernel development. It might also cause boot failures.
> -	  For 64-bit this is recommended on all multiprocessor Opteron systems.
> -	  If the system is EM64T, you should say N unless your system is
> -	  EM64T NUMA.
> +	  For 64-bit this is recommended if the system is Intel Core 7i
> +	  (or later), AMD Opteron, or EM64T NUMA.
> +
> +	  For 32-bit this is only needed on (rare) 32-bit-only platforms
> +	  that support NUMA topologies, such as NUMAQ / Summit, or if you
> +	  boot a 32-bit kernel on a 64-bit NUMA platform.
> +
> +	  Otherwise, you should say N.
>  
>  comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
>  	depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI)

Core i7, not Core 7i.

-- Chris

Re: [PATCH] change CONFIG_NUMA description

[Ingo Molnar]

* Chris Snook <csnook@redhat.com> wrote:

>> +	  For 32-bit this is only needed on (rare) 32-bit-only platforms
>> +	  that support NUMA topologies, such as NUMAQ / Summit, or if you
>> +	  boot a 32-bit kernel on a 64-bit NUMA platform.
>> +
>> +	  Otherwise, you should say N.
>>   comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
>>  	depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI)
>
> Core i7, not Core 7i.

fixed that, thanks!

	Ingo

Re: [PATCH] change CONFIG_NUMA description

[KOSAKI Motohiro]

> > +	  For 64-bit this is recommended if the system is Intel Core 7i
> > +	  (or later), AMD Opteron, or EM64T NUMA.
> > +
> > +	  For 32-bit this is only needed on (rare) 32-bit-only platforms
> > +	  that support NUMA topologies, such as NUMAQ / Summit, or if you
> > +	  boot a 32-bit kernel on a 64-bit NUMA platform.
> > +
> > +	  Otherwise, you should say N.
> >  
> >  comment "NUMA (Summit) requires SMP, 64GB highmem support, ACPI"
> >  	depends on X86_32 && X86_SUMMIT && (!HIGHMEM64G || !ACPI)
> 
> Core i7, not Core 7i.

Good catch.
Tanks.